Shaeel Hashmi

I'm a Computer Science student who builds full-stack web applications, REST APIs, automation tools, and machine learning solutions. I've developed real-time chat, speech-to-text, news recommendation, and notes applications using modern web and Python-based stacks.

At Nexative Technologies, I implemented a PyTorch fine-tuning pipeline for a Hindi speech-to-text model, built a CNN-based image character classifier, and researched multitenant architecture implementation. I also fine-tuned Whisper-tiny for Urdu speech recognition, reducing WER to 59.57% on Common Voice and 51% on Google Fleurs.

At Ninjas Code in Islamabad, I automated issue creation and pull-request merging with Python and Selenium, reducing task time by 30%. I contributed to the open-source Enatega project by opening 40 issues and merging 10 Dependabot pull requests across four repositories each day.

I'm active in competitive programming, placing second among more than 80 university teams at Aurex and earning First Runner-Up at COMPEC. I enjoy creating scalable, efficient web solutions and contributing practical improvements to open-source software.